Documentation for pathology-image-features

Formalization of quantiative image feature extraction from histology images.

Motivation

pathology-image-features was developed to provide an easy-to-use interface for standardization of quantitative image feature extraction.

pathology-image-features uses the *rdflib* library to apply ontology principles to Python objects. Creation of image features in this manner readily prepares quantitative image features for storage in graph databases, which are especially beneficial for distribution of FAIR (Findable, Accessible, Interoperable, and Reusable) datasets. Additionally, improving the ability to navigate different types of image features simplifies the process of translating clinical criteria to quantitative measurements.

While this library was originally designed to meet the needs of the pathology domain, the same principles can be applied to general images as well.

Citation

If you found this project useful in your own work, please consider citing:

Table of contents

• Background
  • Why use ontology? What is ontology?
  • Integrating ontology in Python
  • Ontology in pathology image analysis
• Installation
• Basic Usage
  • Importing Terms as classes
  • Adding Instance attributes
  • Dynamic Subclass creation
  • Data Transformations
• Advanced Usage
  • Properties
  • Restrictions
• Working with Graph Data
  • Serialization
  • SPARQL queries
• Visualization
  • Graphviz Integration
  • Specific Entity Visualization
• I/O Operations
  • Annotations I/O